In the prior art, transistors are often used as high-frequency amplifiers. In applications where the operating frequencies are below microwave frequencies (i.e., &lt;2 GHz), silicon bipolar transistors are typically used. Although silicon bipolar transistors typically do not have performance characteristics as good as field effect transistors ("FETs"), bipolar transistors have the advantage that their associated bias circuitry is generally less complicated than that traditionally needed for FETs. In particular, it is well known how to bias silicon bipolar transistors with only a single voltage source.
In microwave frequency applications, FETs are more commonly used than bipolar transistors because FETs generally have better gain, noise and linearity characteristics than bipolar transistors. Even though FETs generally have better performance characteristics than bipolar transistors, the use of FETs in low frequency applications (i.e., &lt;2 GHz) has not been common for two reasons. First, the performance characteristics of bipolar transistors, while generally not as good as that for FETs, are often satisfactory for many applications. And second, FETs are generally more expensive and difficult to use than bipolar transistors because FETs have required two disparate voltage sources to be properly biased. Although for some applications it is acceptable to "float" a FET (i.e., bias it with only one voltage source), this causes the performance of the FET to be substantially degraded.
FETs are typically biased to function as class-A amplifiers and as such, benefit from a constant drain current and drain to source voltage, regardless of the signal level. But because of manufacturing variations, certain physical characteristics of individual FETs (e.g., pinch-offvoltage and transconductance) can vary and, therefore, additional circuitry may be necessary to compensate for these variations and thus ensure a constant drain current and drain to source voltage. Of course, there are some applications where variations in the physical characteristics of FETs are not a problem and can be safely ignored.
FIG. 1 is a schematic diagram of an amplifier circuit containing a FET and a feedback control system that maintains the drain current and voltage of the FET at specified levels, regardless of variations in the physical characteristics of the transistor. The drain current is sensed in a resistor, compared to a reference, and a negative gate bias is adjusted to yield the desired set of conditions. Some high performance applications use operational amplifiers to perform the feedback function. The fundamental concept, however, is the same. This type of bias circuit works well and is widely used. It's principal disadvantage is that it requires two voltage sources.
FIG. 2 is a schematic diagram a typical circuit that uses the "floated source" method of biasing a FET with only a single voltage source. Current through the source resistor effectively biases the FET gate negative relative to the source. The source is bypassed with a capacitor to ground so as to mitigate against negative feedback. While this technique is generally acceptable for low frequency applications, it is disadvantageous for high frequency applications because the performance of the circuits significantly degrades. Another disadvantage of the "floated source" technique is that variations in the physical characteristics of individual FETs, which occur in normal manufacturing, can result in unacceptable variations in the drain current and drain to source voltage.
Occasionally, these variations are partially compensated for by including in the circuit a variable resistor in series with the source (as shown in FIG. 2). Once the circuit is fabricated, the variable resistor is adjusted to so that the circuit yields the desired drain current. This fix is problematic, however, in that it results in variations in the drain to source voltage, which may be a critical parameter in some applications. Depending on the complexity of the circuitry and the sophistication of the designer, the variable resistor can be fabricated in numerous ways (e.g., a potentiometer, multiple wirebonded parallel resistors on a hybrid circuit, laser trimming of thin-film resistors, etc.). In all of these cases, however, the drain to source voltage remains variable.